EP0418283A1

EP0418283A1 - Antibodies

Info

Publication number: EP0418283A1
Application number: EP89906389A
Authority: EP
Inventors: Agamemnon Antoniou Epenetos; Nigel Stephen Courtenay-Luck; Gregory Byron Sivolapenko
Original assignee: Imperial Cancer Research Technology Ltd
Current assignee: Cancer Research Horizons Ltd
Priority date: 1988-05-23
Filing date: 1989-05-23
Publication date: 1991-03-27
Also published as: GB8812137D0; WO1989011537A1; JPH03504323A; AU3698789A

Abstract

On produit des anticorps anti-idiotypes2 contre un antigène en inoculant des animaux avec un anticorps contre l'antigène et en récupérant les anticorps anti-idiotypes2 à partir de fluides corporels. On peut également utiliser le procédé pour produire des cellules sécrétrices d'anticorps ainsi que des lignées celulaires immortalisées. Les cellules et les anticorps anti-idiotipes2 sont utiles en thérapie et diagnose.Anti-idiotype2 antibodies against an antigen are produced by inoculating animals with an antibody against the antigen and recovering the anti-idiotype2 antibodies from body fluids. The method can also be used to produce antibody-secreting cells as well as immortalized cell lines. Anti-idiotipes2 cells and antibodies are useful in therapy and diagnostics.

Description

ANTIBODIES

2 The present invention relates to anti-idiotype antibodies, to a process for their production and to their use as therapeutic agents and diagnostic reagents.

When an antibody is administered to an animal it is possible that it will act as an immunogen and elicit an immune response such that the animal produces its own antibodies which recognise sites on the original antibody. Some of these will recognise the antigen binding site of the original antibody (i.e. the paratope) whilst others recognise sites elsewhere on the variable region (i.e, idiotopes) . The anti-paratopic antibodies have a binding site which fits the paratope of the original binding site and which.must therefore have the same or very similar structure to the original antigen. These antibodies are know as anti-idiotype antibodies and are regarded as bearing an "internal image" of the antigen.

Anti-idiotype antibodies (hereafter "anti-Id ") can be isolated and administered to other animals in which they will also act as an immunogen and elicit production of

₂ f ur ther antibod ies ( anti- id iotype antibod ies ; hereafter

"anti-Id 2") which recognise the paratope of the anti-Id1 and thus contain an internal image of the original

2 antibodies' paratope. These anti-Id are therefore capable of recognising the original antigen.

In the past it has been suspected that administration of an antigen will lead to a cascade of antibody, anti-Id 1, anti-Id2 etc formation which may play an important role in the regulation of the immune response.

2

However, to date, anti-Id formation has only been achieved experimentally by the process of isolating the antibody, using it as an antigen, isolating the anti-Id and using that in turn as an antigen.

It has now surprisingly been discovered that 2 anti-Id formation can be stimulated directly by adminiεtration of a monoclonal antibody or fragment thereof

2 and that immortalisation of anti-Id producing cells

2 permits the production of monoclonal anti-Id (hereafter

"MA-Id ) which are useful as therapeutic agents and diagnostic reagents.

Accordingly the present invention provides a

2 process for the production of anti-idiotype antibodies or fragments thereof against an antigen which process comprises inoculating an animal with monoclonal antibodies or fragments thereof against the antigen and recovering antibodies against the antigen from a body fluid of the animal.

The animal may be a human or non-human mammal though ethical considerations may limit the use of humans in this process. Of the non-human mammals, conventional laboratory rodents,^" especially rats and mice, and primates will be preferred.

MAb's may be readily available for instance from commercial sources or by culturing deposited hybridomas.

Alternatively they can be produced _ab initio by the Kohler

& Milstein methodology.

Inoculation is by any of the usual routes such as intraperitoneal, intravascular , intramuscular or subcutaneous injection and the quantity of the monoclonal antibody (hereafter "MAb") inoculated will be selected depending upon the immunogenicity of the MAb. Typical quantities are from 1 to lOug for mice and about 30mg for humans. In order to elicit satisfactory antibody production the inoculation may be repeated once or more at intervals of a few days to a few weeks. The skilled person is aware of the standard schedules for such inoculations.

Antibodies may be recovered by conventional separation techniques. Since the inoculated MAb will rapidly be cleared from the body fluids, any antibody found in the body fluids after a suitable interval which is specific for the antigen against which the MAb is directed. will necessarily be an anti-Id2. Fragments of anti-Id2 may be produced by conventional methods. Preferred fragments are F(ab' )- fragments.

The invention also provides a process for producing cells capable of producing anti-Id against an antigen comprising inoculating an animal with MAb against an antigen and recovering cells from a tissue or body fluid of the animal which secrete antibody against the antigen.

Inoculation is as described above. The cells may be recovered for instance from the peritoneal cavity, the spleen or the blood, splenocytes and peripheral blood lymphocytes being particularly convenient.

It has been found that the production by an 2 animal of anti-Id producing cells rises sharply over a few days and that cells obtained at about the mid-point of this rise are most useful. The rise can be detected simply by

₂ daily monitor ing of circulating anti-Id levels .

The invention alternat ively pr ov ides a process

2 for producing anti-Id against an antigen or cells capable

2 of producing anti-Id against an antigen comprising

2 recovering anti-Id or cells from a body fluid or tissue of an animal having immunity to MAb against the antigen.

The invention further provides a process for

2 producing immortal cells capable of secreting anti-Id against an antigen which process comprises immortalising a

2 cell capable of secreting anti-Id against the antigen.

Immortalisation may be effected by for instance treatment with Epstein-Barr virus (EBV) or by fusion with an immortal cell such as a tumour cell especially a myeloma cell. Techniques for achieving such immortalisation are well known to the skilled person.

The invention also provides (i) immortal cell lines capable of secreting anti-Id 2 against an antigen and

2 (ιi) monoclonal anti-Id produced by such cell lines.

The process of the present invention is particularly advantageous in that it permits the binding site ( paratope) of a first MAb to be copied and produced on a large scale in the form of a second MAb . The second MAb may be of a different species to the first ( useful for instance in avoiding spec ies - spec ies interactions and overcoming problems of instabil ity in the hybr idoma cell wh ich secretes the or ig inal MAb) and/or it may be of a d ifferent immunoglobul in class or subclass from that of the f ir st MAb (useful especially with human antibod ies where to date most MAb * s are IgM yet IgG is normally requ ired for therapy and diagnosis) . The process also repre sents a considerable saving of time and labour over previous attempts to generate anti-Id 2 by iterative immunisation, recovery and purification steps.

In particular aspects of the invention the processes therefore involve either the use of MAb derived from a species different from the inoculated animal or the

2 selection and recovery of an anti-Id (or of cells

2 producing an anti-Id ) of a different immunoglobulin class or sub-class to that of the MAb.

2

Anti-Id produced according to the invention

2 especially MA-Id . and fragments thereof , may be used in any conven tional appl ication in therapy - for instance for pass ive immunisation , or for tumour destr uction using an i munotoxin and in any conventional diagnostic application s uch as Rad io I munoassay , Enzyme Linked Immuno- sorbent

Assay and similar assay procedures .

The invent ion therefore provides 2 a) Anti-Id for use in a method of therapy or d iagnos is practised on the human or animal body .

₂ b) Cells capable of secreting anti-I for use in a method of therapy or diagnosis practised on the human or animal body. c) The use of anti-Id² or cells capable of secreting anti-Id in the preparation of a medicament for use in a method of therapy or diagnosis practised on the human or animal body. -5-

d) A method of therapy or diagnosis comprising administering to a human or non-human animal an effective, non-toxic amount of anti-Id 2 or cells capable of secreting

2 anti-Id . e) A diagnostic assay procedure comprising contacting a sample suspected to contain an immunogen with anti-Id².

2 f) A diagnostic assay kit comprising anti-Id

2 or cells capable of secreting an anti-Id .

2

In the above (a) to (f) the anti-Id and cells

2 capable of secreting anti-Id are preferably produced according to the processes described above.

The invention will now be illustrated with reference to the figures on the accompanying drawings in which:

Figure 1_ binding of patient's serum antibodies to the administered murine Mab (HMFGl) and a control antibody of the same isotype, but idiotypically unrelated (1141), in an ELISA, after a single therapeutic administration of radiolabeled murine Mab.

Figure 2_ shows typical comparative binding of patient's serum antibodies, after 2 or more therapeutic administrations of radiolabeled murine Mabs, to the administered (HMFGl) and control (1141) murine monoclonal antibodies coated onto 96-well microtiter plates. The anti-idiotypic component is obtained by subtraction of binding results for 1141 from HMFGl. Nen-specific binding to antigen-free wells was evaluated by using the ELISA coating buffer (E.C.B.) alone.

Figure 3_^ shows the results obtained by comparing the binding of serum antibodies from patient 11 to F(ab')₂ fragments of the administered (HMFGl) and control (1141) monoclonal antibodies with the anti—idiotypic response being obtained by subtraction. Essentially identical results were obtained when sera from other patients who had received two or more therapeutic administrations of radiolabeled murine Mabs.

Figure _4 shows the percentage that the administered murine monoclonal antibody HMFGl (10y_/g/ml) is inhibited by patient 15's pre and post 3rd therapy serum (1/10 dilution).

Figure 5_ shows the inhibition of the administered murine monoclonal antibody binding to its antigen by human antibodies present in the serum of patient 15 prior to (pre) and post four therapeutic adminstrations of radiolabeled murine monoclonal antibodies (PI, P2, P3 and P4). -7-

Figure 6_^ shows the binding of the administered murine monoclonal antibody, HMFGl at lO g/ml (A) to its antigen (MFG) and the inhibition of this binding by patient 15's pre-therapy serum (B) post 3rd therapy serum (C) post 3rd therapy Ig devoid serum (D) and post 3rd therapy serum Ig fraction, after protein-A purification of post-third therapy serum.

Figure 1_ shows the binding of patient 15's (Table 1) serum antibodies to the tumor associated antigen (that used to raise the administered murine Mab) prior to (-{7^")-) and post 3rd therapy, (-^>-)- Also shown is the effect on the anti-tumor response of removing serum antibodies reacting with the administered murine Mab (-^ —) including the anti-id' antibodies.

Figure 8_ shows the results of autoantibody screening by immunofluorescence. Fig. 8A shows that no pre-existing autoreactive antibodies were detectable in this present patient (patient 15) when assayed on fresh frozen sections of rat liver. No other patients' pre-therapy serum gave positive staining (Table 1). After receiving three intraperitoneal administrations of monoclonal antibodies, patient 15's serum contained high levels of antibodies directed against connective tissue components of liver (8B), kidney glomerulus (8D) and diaphragm (8E). This generation of autoantibodies was no longer detectable after a fourth administration, as shown in Figure 8C, indicating autoimmune regulation.

Figure 9_ illustrates, diagramatically, the findings of this study. These are, that the adminstration of murine Mab directed against tumor associated antigens, results in the generation of an anti-mouse Ig response. Some antibodies produced in this response are against constant domain determinants of the murine Ig (1) while others are against variable domain determinants, i.e. anti-idiotypic, ₍2₎ (anti-id¹). Either these anti-id¹ antibodies, ⁽in particular those that are anti-para-topic⁾ or tumor associated antigen shed from targeted cells, give rise to the generation of antibodies having binding specificities similar to that of the administered murine Mab ⁽3⁾ i.e. anti-id². In some patients, autoreactive antibodies are also generated, either, as a direct result of administering murine Ig (4b) or in response to antigenic determinants on the anti- urine Ig antibodies, ⁽4a⁾.

Figure 10: Typical antibody response of a rat (Rat 4), against the HMFG2 that was administered, and against an irrelevant mouse monoclonal antibody (11.4.1) that is of the same class and subclass with HMFG2.

The antibody titres are measured at serum dilution 1:10; and at various time points (before, after the 1st and after the 2nd immunisation).

Figure 11: The rat anti-HMFG2 and anti-11.4.1 response represented as area underneath the curve as shown in Figure 1.

* Only two serum samples available, before and after the 1st immunisation.

+ Rats received the first two injections with Freud's ^•adjuvant. Figure 12: Immunoperoxidase staining using the GMN-B4 antibody (supernatant).

Figure 13: Negative immunoperoxidase staining of the same as in Figure 3 section using supernatant of an irrelevant rat monoclonal antibody.

Figure 1 : Immunoperoxidase staining of the same section as in Figures 3,4 using the HMFG2 antibody (purified).

Figure 15: Immunoperoxidase staining of a section using the GMN-B4 antibody (supernatant).

Figure 16: Negative immunoperoxidase staining of the same section as in Figure 6, using an irrelevant rat monoclonal antibody (supernatant).

Figure 17: Immunoperoxidase staining of the same section as in Figures 6, 7 using the HMFG2 antibody (purified).

Cells capable of secreting anti-idiotype antibodies may also be produced by reco binant DNA techniques. For this, nucleic acid (e.g. chromosomal DNA or messenger RNA) encoding the anti-Id antibodies is recovered by conventional techniques including amplification by the polymerase chain reaction (PCR) from selected antibody-secreting cells of a human or animal inoculated with antibody against the target antigen. Appropriate expression vectors containing the coding sequences in open frame register with any necessary regulatory sequences such as promoters, initiation and termination signals are produced by conventional methods and are transfected into suitable host cells to form an expression system. The host cells must be capable of expressing the coding sequences under appropriate culture conditions and will preferably (but not essentially) secrete the anti-Id² antibodies so produced into the culture medium from which they may be recovered by conventional techniques.

Partially when it is desired to produce human anti-Id² antibodies by this method, the human will be inoculated with an antibody as part of a therapeutic treatment. Cells such as peripheral B lymphocytes will then be recovered from the blood and selected by panning using antigen bound to a solid support. Cells expressing the anti-Id² or their surface will bind to the antigen, and are the source of the nucleic acid. Amplified copies of the coding sequences are obtained using the PCR primed with nucleolide fragments having a sequence complementary to human immunoglobulin sequences and such amplified copies are used to produce the expression vectors. EXAMPLE 1_

The invention will now be illustrated by the following Examples which are not intended to limit the scope of the invention in any way:

The development of the hybridoma technique (1) has led to the application of monoclonal antibodies (Mabs) in clinical diagnosis of neoplasia (2-7). More recently Mabs coupled to radioactive isotopes have been used for the therapy of some cancers (8), with increased tumor localisation being achieved by regional ^'administration (9-12).

The immunological consequences of ' the jj vivo use of murine antibodies have been studied by a number of groups. It has been shown that some, but not all patients, develop human anti-murine immunoglobulin (Ig) antibodies (13-17), and that some patients develop anti-idiotypic antibodies (18-21) - that i^*s, antibodies which bind to the variable region of the administered Mab. These anti-idiotypic antibodies (anti-Id¹) will contain a mixture of specificities, some recognising the antigen binding site itself (paratope) and others recognising sites elsewhere on the variable region (idiotopes). The anti-paratopic antibodies will -mimic the original antigen, thus providing an "internal image" of that antigen. Anti-Id¹ antibodies can themselves act as antigen, leading to an anti-Id² response; at least some of these anti-Id² antibodies (those that are anti-paratopic) will have specificity for the original antigen. The generation and maintenance of such an "idiotypic network" is thought to play an important part in immune regulation (22). Pertubation of this network have been shown to lead to regulatory changes in the immune system's response to self, antigens (23-28).

We have analysed the specificity of serum antibodies in patients with ovarian cancer treated with repeated intra-peritonea therapeutic doses of radiolabeled murine Mabs. We demonstrate the generation of anti-id¹ antibodies, anti-Id² antibodies with specificity for the tumor associated antigen and transient autoantibodies to connective tissue components of many organs.

Materials and Methods

1. Patients

Fifteen patients with stage III or IV histologically proven ovarian cancer gave written informed consent to enter a phase III trial of ¹³¹Iodine-labeled murine Mab administered intraperitoneally for the treatment of advanced ovarian cancer.

2. Antibody protocols

Patients received between 5 and 20mg of Mab labeled with between 50 and 150 mCi (1.85 - 5.55 GBq) of ¹³¹ Iodine, injected intraperitoneally. Patients receiving fractionated therapy, were given an initial dose of 100 mCi (3.7 GBq) ¹³¹iodine labeled to lOmg of Mab, followed by two or three 50 mCi (5-10mg antibody) doses at intervals of one month between administrations (Table 1).

3. Monoclonal antibodies

HMFGl and HMFG2 are both murine IgGl antibodies which bind to a large mucin like molecule normally produced by the lactating breast, but also expressed by the majority (>90%) of ovarian, breast and other carcinomas of epithelial origin (29). AUAl is a mouse IgGl antibody which detects an antigen expressed by a wide range of adeno-carcinomas, including approximately 75% of carcinomas of the ovary (30). H17E2- is a mouse IgGl antibody directed against placental alkaline phosphatase. This enzyme is expressed as a surface membrane antigen on many neoplasms, including 60-85% of ovarian carcinomas (31). 11.41 is a mouse IgGl antibody raised against a murine la antigen (32).

4. ■ ELISA system to determine the presence of anti-idiotypic antibodies

In order to demonstrate anti-idiotypic antibodies, sera were assayed in parallel against the administered Mab and against an idiotypically unrelated Mab of the same isotype (control antibody) The wells of rows 1-6 of an ELISA plate were coated with the administered Mab (500ng/well) and rows 7-12 with the control antibody (500ng/well) . Both antibodies were diluted in bicarbonate buffer pH9.6. The use of a single plate enabled a comparison to be. made under identical experimental conditions. Details of the method have been described previously (15) .

5. Inhibition assay to demonstrate anti-paratropic antibodies within the anti-idiotypic response

Patient's sera that showed anti-idiotypic activity in the above ELISA were pre-incubated with the administered murine Mab (lOμg/ml) at various dilutions of the sera. Incubation was for 18 hr at 4°C. The ability of patient's sera to inhibit the binding of the administered murine Mab to its target antigen was then assayed by ELISA using purified milk fat globule (MFG) antigen as target (3yg/ml - 30ng per well). This was the antigen to which the administered murine Mab was raised. Equivalent concentrations of the administered Mab were incubated for the same period in PBS. and used in the assay as a positive control.

As negative control, the patient's pre-therapy se um alone was used.

6. Purification of the immunoglobulin component from whole serum by protein A affinity chromatography

To ensure that inhibition of binding of the administered murine Mab to its antigen was due to anti-idiotypic antibody and not to free circulating antigen, the Ig fraction from a sample of patient 15's post 4th therapy serum was purified by protein A chromatography as follows. Serum was dialysed for 48 hr at 4°C against 0.1M phosphate buffer, pH8.0. After dialysis, the serum was incubated overnight at 4°C with protein A-linked sepha.rose 4B beads, using continuous rotation. Next day, the beads were poured into a column and the Ig depleted serum stored at 4°C for future testing. The protein A column was equilibrated using O.lM phosphate buffer; this buffer was then ^"used to elute any unbound or non-specifically bound protein. When the absorbance (280nm) of the eluate was zero, the Ig fraction was eluted off the beads using O.lM citrate buffer at pH6. After protein A chromatography the whole serum, the Ig devoid serum, and the affinity purified Ig fraction were assayed for a) ability to bind to murine lg b) ability to inhibit the binding of the Mab to its target antigen c) molecular content using 7.5% non-reduced NaDodSO4-PAGE.

7. Anti-Id² (anti-tumour) antibody assay

Prior to assaying for the presence of anti-tumour (anti-Id²) antibodies, the anti- murine Ig and anti-Id¹ antibodies were removed from patient's sera using an HMFGl (administered Mab) affinity column as previously described (18,19), and the efficiency of removal checked by ELISA. After total depletion of anti- urine Ig and anti-Id¹ antibodies, sera were then assayed for their anti-Id² content by ELISA using microtitre plates coated with MFG antigen (30ng/well). As positive control we used the original administered murine Mab (with specificity for the tumour antigen) and as negative control the patient's pre-therapy serum was used. The pre-therapy serum enabled assessment of pre-existing cross-reative antibodies to the MFG antigen.

8. Autoantibody assay

Autoantibodies were assayed by immunofluorescence on fresh frozen tissue sections of rat liver, kidney, diaphragm and stomach. Sections were air dried, fixed in acetone and stored at -20°C. Just prior to use, sections were thawed, washed in PBS and incubated for lhr at room temperature with 50μl of patient's serum at dilutions of 1/40, 1/80, 1/160, 1/320. Following this, sections were washed for 30 min in PBS, incubated for 15 min with a 1/40 dilution of fluorescein conjugated sheep anti-human IgG antibody (Wellcome Diagnostics, Dartford, England), then washed as before, mounted in hydromount (National Diagnostics, Somerville, NJ) and screened_. using a Leitz UV microscope equipped with epi-illumination optics. For controls, patient's serum was replaced by PBS, serum from healthy controls, serum from patients with neoplastic conditions identical to those under study but receiving no Mab therapy, and lastly pre-therapy serum from patients participating in this study. In order to test .for the possibility of false positives due to anti-mouse Ig antibodies in the patient's serum binding to rat Ig in the tissue sections, sections were incubated with fluorescein-labeled rabbit anti-rat Ig antibody. This enabled us to determine the distribution of endogenous rat Ig in the tissues. No binding of patient's sera to this endogenous rat Ig was detected. RESULTS

1. Assay for anti-idiotypic antibodies in patients serum post first and subsequent therapy

No anti-Id¹ antibodies were detected after a single administration of Mab, (Fig.l). However, after two (patients 11, 12 and 15) or three (patients 13 and 14) therapeutic administrations, elevated binding of human anti-murine Ig antibodies, to the administered antibody over that to the control antibody was observed (Table 1, Fig.(2), indicating the development of anti-Id antibodies.

When microtiter plates were coated with F(ab') fragments, in order to reduce the number of constant region antigenic determinants of the murine Mabs, the preferential binding to the administered antibody was enhanced (as shown in Fig.3 for patient 11). This finding was also true for other patients showing an anti-Id¹ response.

4. Inhibition of the binding of administered murine monoclonal antibody to its antigen by patient's serum

Further confirmation of the presence of anti- idiotypic antibodies was obtained from inhibition studies. Since the binding site of some anti-Id¹ antibodies will mimic the original antigen (providing an internal image of that antigen) such antibodies should inhibit the binding of the administered murine Mab to its original tumour antigen (MFG) . The results of these inhibition

TABLE 1

Patient Administered Dose Presence Presence Prese Monoclonal Administered of anti-Id of anti-Id of au Antibody antibody antibodies antib in serum in serum in se

1 HMFG2 10mg - - —

2 HMFG2 lOmg - - -

3 HMGG2/H17E2 lOmg - - -

4 HMFG2 lOmg - - -

5 HI7E2 lOmg - - -

6 H17E2 lOmg - - _

7 HMFG2 17mg - - -

8 HMFGl 20mg - - -

9 AUA1 20mg - - -

10 HMFGl lOmg — — — HMFG2 lOmg

11 HMFG2 7mg HMFG2 6mg + + +

12 HMFGl 12mg - HMFG1/HI7E2 lOmg + +

13 HMFGl 2mg HMFGl lOmg HMFGl 12mg + + + 4 HMFG2 2mg HMFG2 lOmg HMFG2 lOmg + + + 5 HMFGl lOmg HMFGl 15mg + ^' + + HMFGl lOmg + + + HMFGl 8mg + +

experiments (Fig.4) showed that prior incubation with either pre- or post first therapy serum only inhibited the binding of the murine Mab to its MFG target antigen by between 8% (minimum) and 25% (Maximum). However, when patient 15's post 3rd therapy serum was used, binding of the murine Mab to MFG antigen was totally inhibited.

Since patient 15 had received 4 therapeutic administrations (Table 1) further, sequential studies were carried out on her serum prior to and after each therapy. The inhibitory (anti-Id² ) activity was found to increase with the number of administrations (Fig.5).

5. Is the inhibition of binding of administered murine monoclonal to its antigen due to genuine anti-Id¹ antibodies or to circulating tumor antigen?

The inhibition assay was repeated using the affinity purified human Ig, the Ig devoid fraction and the unseparated serum from patient 15 post fourth therapy (Fig. 6). Both unfractionated serum (C) and the purified Ig fraction (E) inhibited the adminstered Mab from binding to its MFG antigen. In contrast the Ig devoid fraction (D) only inhibited by a level comparable to that shown by pre—therapy serum (B). Hence the inhibitory activity is due to antibody (anti-Id¹) and not to circulating antigen.

6. Generation of anti-tumor (anti-Id²) antibodies

To study the possible development of anti-tumor (anti-Id² ) antibodies we tested anti-Id¹ positive sera for increased binding to^' the MFG tumor associated antigen. Of all five patients tested, serum from patient 15 showed the greatest binding. Depletion of the anti-murine Ig activity from these sera (using an anti-mouse Ig affinity column) did not result in any significant drop in the anti-tumour activity (Fig.7). Hence the anti-tumor activity is not due to cross-reactivity of antibodies whose primary specificity i s for murine Ig .

8. Generation of antibodies reacting with self antigens

Serum samples from four of the five patients positive for anti-idiotypic antibodies were positive in the frozen section auto-antibody assay (see Table 1), showing binding to connective tissue components of liver (Fig 8B). kidney glomerulus (Fig 8D) and diaphragm (Fig 8E). Our sequential study of patient 15 showed that pre-therapy, post first therapy and post second therapy sera were all negative. In contrast post third therapy serum gave intense staining (Figure B). However this patient's post fourth therapy serum was virtually negative again (Figure 8C). Thus development of auto antibodies in patient 15 was transient. No other patient received further therapy after the development of autoantibodies.

DISCUSSION

The clinical application of murine Mabs has raised several issues, one of the most crucial being the problem of sensitization of the recipient to the administered xenogeneic protein (13-17). Repeated j^n vivo use of such proteins is not viable; immunization of patients would cause immune complex disease and also abrogate any therapeutic effect due to rapid immune clearance of administered antibody.

We analysed the sequential development of human anti-mouse Ig responses in patients receiving murine Mabs therapeutically for ovarian cancer. We previously described the generation of human antibodies to the Fc and F(ab')₂ regions of murine Ig (15). Here we extend our data by defining the anti-idiotypic component to this response.

We demonstrate that patients develop anti-idiotypic antibodies, identified as antibodies that bind only to the administered Mab. These could have specificity for the antigen binding site on the therapeutic antibody (anti-paratope) ; alternatively, they might detect determinants elsewhere on the variable domains (anti-idiotope) . Anti-Id¹ antibodies that bind to the paratope on the administered murine anti-tumor Mab will mimic the tumor antigen and hence will form an internal .image of that antigen. We therefore studied the ability of patient's anti-idiotypic antibodies to replace antigen in inhibition studies. We show that the purified human Ig fraction from serum of patients receiving two or more administrations of murine Mab totally blocks the binding of the therapeutic antibdoy to its target tumour antigen. Thus, anti-Id¹ antibodies are produced by these patients. The use of the purified human Ig fraction for these studies excluded the possibiltiy that inhibition was due to soluble tumor antigen [known to be present in the serum (33)]. This may have been a problem in other studies where the anti-Id¹ antibodies could have been co-purified with the tumor antigen.

It has been suggested that the use of human Mabs, or of genetically engineered "humanized" antibodies where only the hypervariable regions are of mouse orgin, might be less immunogenic than xenogeneic antibodies. However, in the light of our data we would predict that these reagents would still generate an anti-idiotypic response. The use of human or chimeric antibodies might provide only a transient therapeutic advantage over existing murine Mabs. Anti-idiotypic (anti-Id¹ ) antibodies would be the limiting factor in the use of antibodies, regardless of their construct.

We investigated the possibility that anti-Id¹ antibodies could act as antigen and generate an anti-anti-idiotypic (anti-Id²) antibody which, by definition, would be an anti-tumor antibody. We have shown elevated anti-tumor activity in the sera of treated patients, as compared with their pre-therapy serum. Although this is consistent with the generation of an idiotypic network, an alternative explanation is that these antibodies are generated in response to shed tumor antigen. This is unlikely for the following reasons. Firstly, the development of a detectable anti-Id² response is always preceded by the generation of anti-Id¹ antibodies. Secondly, pre-therapy serum contains only low levels of anti-tumor antibody activity, despite the fact that there may be high levels of circulating tumor antigen. This is in contrast to the elevation in anti-tumor activity that is seen after Mab therapy.

The generation of anti-tumour antibodies via the idiotypic network could be of benefit tα patients by up-regulating their immune response to their own tumor (18). Reports of tumor regression occurring long after clearance of mouse Ig suggests that anti-idiotypic antibodies may have acted as therapeutic agents (34). The discovery of such antibodies has stimulated the concept of making idiotypic vaccines, and animal studies using tumor specific idiotype vaccines have already generated encouraging results (35-36). Anti-Id¹ antibodies used as vaccines could stimulate the in vivo production of anti-tumor antibodies. However, the problem of using xenogeneic proteins would still be present. To overcome this, one could isolate B-cells from patients exhibiting an anti-Id¹ response and immortalise them by EBV-transformation. The EBV-transformed cell line would then be used to provide human monoclonal anti-Id¹ antibody in large amounts.

Alternatively, the generation of anti-tumour antibodies might be achieved by immunization with purified tumor antigen. The lack of a good anti-tumor response in patients with cancer indicates that tumor antigen is poorly immunogenic. However, recent advances in molecular biological techniques have permitted the isolation of the gene encoding the protein moiety of the HMFGl antigen, and in the future appropriate alteration of this gene may yield an optimally immunogenic vaccine (37).

Finally, in the generation of an idiotypic network not all antibodies will be anti-paratopic. Those that are not, may cross react with self antigens in the host. We found such autoreactivity in four of the patients who had received two or more doses of Mab therapy, with reactivity against self antigens located on connective tissues. One of these four patients (no. 15) received a further dose of Mab therapy. Her serum no longer showed significant binding to autoantigens suggesting that these autoantiboides are transient, and that such autoimmunity will not present serious clinical problems. Regulation is probably the result of anti-idiotypic control, as has been observed in other autoimmune conditions (38), we are however unable to draw any firm conclusions because only four patients developed autoantibodies. Indepth studies are required to dete.rmine the nature of the antigens defined by these autoantibodies and to identify the mechanism responsible for their suppression.

In conclusion, although the intraperitoneal route of administration of therapeutic monoclonal antibodies has the dual advantage of a high targetting efficiency and a low risk of immediate hypersensity due to circulating immune complexes (serum sickness), it may have one disadvantage: the peritoneum is an efficient site for generating an immune response - due to the high content of antigen-presenting macrophages, leading to anti-mouse Ig responses directed against both idiotypic and non-idiotypic determinants.

However, the effects of generating an anti-idiotypic response are not necessarily detrimental. It may be possible to turn such antibodies to the patient's advantage once we know more about the mechanisms controlling the idiotypic network and can tip the balance in favour of the anti-Id² generation - thus providing the patient with auto-anti—tumor antibodies. Reference for Example 1

1. Kohler G. and Milstein C. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 1975, 256, 495-497.

2. Mach J.P., Buchegger F., Forni M. et al. Use of radiolabelled monoclonal anti-CEA antibodies for the detection of human carcinomas by external photoscanning and tomoscintigraphy. Immunology Today 1981, 1, 239-249.

3. Farrands P.A., Perkins A.C., Pimm M.V. , Hardy J.D., Embleton M.J., Baldwin R.W. and Hardcastle J.D. Radioimmunodetection of human colorectal cancers by an anti-tumour monoclonal antibody. The Lancet 1982, ii, 397-400.

4. Epenetos A.A. , Britton K.E., Mather S. et al Tar ettin of lodine-123-labelled tumour associated monoclonal antibodies to ovarian, breast and gastrointestinal tumours. Lancet 1982, ii, 999-1005.

5 Epenetos A.A. , Snook D., Hooker G. et al. Indium-III labelled monoclonal antibodies to placentalalkaline phosphatase in the detection of neoplasm of testis, ovary and cervix. Lancet 1985, ii, 350-35-3.

6. Kemshead J.T., Goldman A., Pritschg J., Malpas J.S., Pritchard J. Use of panels of monoclonal antibodies in the differential diagnosis of neuroblastoma and lymphoblastic disorders Lancet 1983, i, 12-15.

7. Larson S.M., Carrasquilo J.A., Krohn K.A., Brown J.P., McGuffin R.W., Perens J.M., Graham M.M. , Hill L.D., -24-

Beamier P.L., Hellstrom K.E. and Hellstrom I. Localisation of 131 I-labelled P97 specific Fab fragments in human melanoma as a basis for radiotherapy. J.Clin. Invest. 1983. 72, 2101-2114.

8. Carrasquillo J.A., Krohn K.A. , Beamier P., McGuffin R.W. , Brown J.P., Hellstrom K.E., Hellstrom I. and Larson S.M. Diagnosis of and therapy for solid tumours with radiolabelled antibodies and immune fragments. Cancer Treatment Reports 1984, 6B, 317-328.

9. Epenetos A.A., Courtenay-Luck N.S., Hainan K.E. et al., Hammersmith Oncology Group and the Imperial Cancer Research Fund: Antibody guided irradiation of malignant lesions: three cases illustrating a new method of treatment. Lancet 1984, i, 1441-1444.

10. Epenetos A.A. , Courtenay-Luck^'N.S. , Pickering D., Hooker G., Durbin H., Lavender J.P. and McKenzie C.G. Antibody guided irradiation of brain glioma by arterial infusion of radioactive monoclonal antibody againster epidermal growth factor receptor and blood group A antigen. Br.Med.Journal. 1985, 290, 1463-1466.

11. Pectasides D., Stewart S., Courtenay-Luck N.S. Antibody guided irradiation of malignant pleural and pericardial effusions. Br.J.Cancer 1986 53, 727-732.

12. Ashorn R. , Ashorn P., Punnonen R. et al. The use of radiolabelled monoclonal antibodies to human milk fat globule membrane antigens in antibody-guided tumour-imaging, and administration of therapeutic dose of labelled antibody in wide spread ovarian carcinoma. Annales Chirugiae urgiae et Gynaecologiae 1985, 74, Suppl. 197: 5-10. 13. Shawler D.L., Bartholomew R.M., Smith L.M. and Dillman R.O. Human Immune response to multiple injections of murine monoclonal IgGl. J. Immunology 1985, 135, 1530-1535.

14. Miller R.A., Oseroff A.R., Stratte P.T. and Levy R. Monoclonal antibody therapeutic trials in seven patients with T cell lymphoma. Blood 1983, 62, 988-995.

15. Courtenay-Luck N.S., Epenetos A.A., Moore R. , Larche, M., Pectasides D., Dhokia B. and Ritter M.A. Development of primary and secondary immune responses to mouse monoclonal antibodies used in the diagnosis and therapy of malignant neoplasms«• Cancer Research 1986, 46, 6489-6493.

16. Schroff R.W., Foon K.A., Beatty S.M., Oldham R.K. and Morgan Jr. A.C. Human anti-murine immunoglobulin responses in patients receiving monoclonal antibody therapy. Cancer Research 1985, 45, 879-885.

17. Oldham R.K., Foon K.A., Morgan A.C, Woodhouse C.S., Schroff R.W., Abrams P.G., Per M. , Schronberger C.S., Farrell M., Kimball E. and Sherwin S.A. Monoclonal antibody therapy of malignant melanoma: iri vivo localisation in cutaneous metastasis after intravenous administration. J. Clin. Oncol. 1984, 2, 1235-1244.

18. Koprowski H., Herlyn D., Lubeck M. , DeFreiras E. and Sears H.F. Human anti-idiotype antibodies in cancer patients: Is the modulation of the immune response beneficial for the patient? Proc.Natl.Acad.Sci.U.S.A. 1984, 811, 216-219.

19. Herlyn D., Lubeck., Sears H. and Koprowski H. Specific detection of anti-idiotypic immune responses in cancer patients treated with murine monoclonal antibody. Journal of Immunological Methods 1985, 85, 27-38.

20. Chatenoud L, Baudrihaye M.F., Chikoff N. , Kieis H., Goldstein F. and Bach J-F. Restriction of the human in vivo immune response against the mouse monoclonal antibody OKT3. The Journal of Immunology 1986, 137, 830-838.

21. Herlyn D., Ross A.H. and Koprowski H. Anti-idiotypic antibodies bear the internal image of a human tumour antigen. Science 1986, 232, 100-102.

22. Jerne N.K. Towards a network theory of the immune system. Ann. Immunol. (Paris) 1974, 125C, 373-389.

23. Plotz N.K. Autoantibodies are anti-idiotype antibodies to antiviral antibodies. Lancet 1983, ii, 824-826.

24. Paul W.E. and Bona C. Regulatory idiotypes and immune networks: a hypothesis. Immunology Today 1982, 9, (3), 230-234.

25. Zanetti M., DeBaels M. and Rogers J. High dregree of idiotypic cross reactivity among murine monoclonal antibodies to thyroglobulin. The J. Immunol. 1983, 131, 2452-2457.

26. Zanetti M., Rogers J. and Katz H. Induction of autoreactivity by anti-idiotypic antibodies.

■ Regulation of the Immune System. 1984, 893-907.

27. Harkiss G.D., Hendrie F. and Nuki G. Cross-reactive idiotypes in Anti-DNA antibodies of system lupus Erythematosies (?) Patients. Clin. Immunol. I munopathol. 1986. 39, 421-430. 28. Essani K., Spinivassappa J., McClintock P.R., Prabhakar B.S. and Notkins A.L. Multiple organ- reactive IgGl antibody-induced by an anti-idiotypic antibody to a human monoclonal IgM auto-antibody. J.Exp. Med. 1986, 163, 1355-1360.

29. Burchell J., Durbin H. and Taylor-Papadimitriou J. Complexity of expression of antigenic determinants recognised by monoclonal antibodies HMFGl and HMFG2 in normal and malignant human mammary epithelial cell. J. Immunol. 1963, 131, 508-513.

30. Spurr N.M., Durbin H., Sheer D., Parkar M., Bobrow L. and Bodmer W.F. Characterisation and chromosomal assignment of a human cell surface 'antigen defined by the monoclonal antibody AUAl. Int. J. Cancer 1980, 38, 831-836.

31. Travers P. and Bodmer W.F. Preparation and characterisation of monoclonal antibodies against placental alkaline phosphates and other human trophoblast associated determinants. Int..J. Cancer 1984, 633-641.

32. Oi V.T., Jones P.P., Gooding J.W., Herzenberg L.A. Properties of monoclonal antibodies to mouse Ig allotypes H-2 and la antigens. Curr.Topic Microbiol. Immunol. 81, 115-119.

33. Canney P.A., Moore M., Wilkinson P.M. and Jones R.D. Ovarian cancer antigen CA125: a prospective clinical assessment of its role as a tumour marker. Brit. J. Cancer 50, 765-769.

34. Raychaudhuri S., Saeki Y. , Fuji H. and Kohler H. Tumour specific idiotype vaccines. J. Immunol. 137, 1743-1749. 35. Dunn P.L., Johnson CA. , Styles J.M., Pease S.S. and Dean C.J. Vaccination with syngeneic monoclonal anti-idiotype protects against a tumour challenge. Immunology 60, 181-186.

36. Spitler L.E., Rio M.D., Khentigan A., Wedel N.I., Briphy N.A., Miller L.L., Harkonen W.S., Rosendorf L.L, Lee H.M, Mischak R.P., Kawahata R.T., Stoude ire J.B., Fradkin L.B., Bautista E.E. and Scannon P.J.. Therapy of patients with malignant melanoma using a monoclonal anti-melanoma antibody-ricin A chain immunotoxin. Cancer Res, 47 1717-1723.

37. Swallow D.M. , Gendler S., Griffiths B., Corney G. , Taylor Papadimitriou and Bramwell M. The human tumour associated epithelial mucins are coded by an expressed hypervariable gene locus PUM. Nature 328. 82-84.

38. Sultan Y., Rossi F. and Kazatchkine M.D. Recovery from anti-VIII: C (antihemophilic factor) autoimmune disease is dependent on generation of anti-idiotypes

. against anti VIII: C autoantibodies. Proc.Natl. cad Sci U.S.A. 84, 828-831.

EXAMPLE 2

Materials and Methods

Immunogen-Cell line

The HMFG2 monoclonal antibody was used as the immunogen. This mouse IgGl recognises a mucin expressed on the T47D adenocarcinoma cell line, as well as on many ovarian, breast and colorectoral carcinomas.

Animals

Six Lou/c rats 6 to 8 weeks old have been used.

Immunisation Schedule:

Half the animals were injected i.p. with 200 μg HMFG2 in PBS, whereas the other half received the same amount in Complete Freund's Adjuvant. Seven days later the same procedure was repeated, the only difference being, that this time the adjuvant was incomplete Freud's. For the following three weeks all the animals received 200 _//g/week HMFG2 in PBS i.v. (tail vein).

Rat anti-mouse immunoglobulin response:

Before and at regular intervals during the immunisation, all the rats were bled, and the sera was tested for the presence of anti-HMFG2 antibodies. The method followed was an already established ELISA. Very briefly, 96-well microtiter plates were coated with HMFG2 and an irrelevant mouse IgGl (11.4.1, which recognises the mouse (H-2K). The test sera was then applied, and after 2hr incubation, a species specific anti-rat Ig peroxidase conjugated, was added. The presence of the positive sera was indicated using ABTS as substrate.

The data is presented as area underneath the curve made of the rats response before and after the 1st and 2nd immunisation (serum dilution 1:10).

Rat anti-tumour response:

All. rat sera were tested for the presence of anti-T47D antibodies, before and during the immunisation with HMFG2. Live cells were incubated with each serum, diluted 1:10 at 4^CC for lhr. The cells were.washed and incubated with a-rat Ig, iodine-125 labelled. The radioactivity was measured using a γ-counter.

Moreover, the same sera were tested against the purified mucin, using an ELISA system. The antigen was coated on a 96-well plate, the sera were incubated for 2 hr, and after the plate was washed, anti-rat Ig peroxidase conjugated was added as a second layer. The substrate used was again the ABTS.

Fusion:

Four days after the last immunisation the best animal from each group was sacrificed, and half the spleen cells were fused with the mouse NSO myeloma cell line. whereas the other half with the rat Y3 myeloma line. The cells were plated out in 96-well microtiter plates, and when they were fully grown, their supernatant was tested for antibody activity.

Screening Methods

Anti-idl: Tissue culture supernatant was incubated with T47D cells at 4°C in the presence of HMFG2. Species specific anti-mouse Ig, iodine 125 labelled, was then applied, and the radioactivity was counted using a r-counter. The HMFG2 incubated with supernatant of an irrelevant rat monoclonal antibody, served as our positive control. Every test supernatant that produced lower counts than the positive control, suggested that it inhibited the HMFG2 and that it may contain rat anti-idl antibodies.

Anti-id2: In order to examine whether the above inhibition was in fact competition (suggesting the presence of rat HMFG2-like antibodies), the supernatant was further tested on T47D cells, using anti-rat Ig, iodine-125 labelled.

Anti-MFG membrane extract activity: The purified mucin was coated on a 96-well microtiter plate, and an ELISA was performed. All positive to T47D supernatantε were tested using the same method as described above (see rat anti-tumour response).

Immunohistochemistry: Paraffin embedded neoplastic tissues that were positive for HMFG2 antibody, were also stained with positive εupernatantε using a well established immunoperoxidase technique.

Characterisation of the anti-id2 antibody: A fraction of the supernatant was purified using a Protein G affinity Chromatography column. We found the class and subclass of the purified rat monoclonal antibody using the BioRad subclass identification Kit.

Results

Rat anti-mouse immunoglobulin response:

Figure 10 shows a typical response (Rat4) against the HMFGl, before and after the 1st and 2nd administration. The same antisera recognise to a lesser extend another antibody (11.4.1), that only has the constant region in common with HMFG2, indicating that the difference may be due to anti-idiotypic response.

The area under the curve was plotted for each rat, the same way as above, and is shown in Figure 11. The difference between the response against the HMFG2 and the 11.4.1. indicates the presence of antibodies that recognise the hypervariable region of the HMFG2.

Rat anti-tumour response:

Table 2 shows the results of the Radioimmun'oassay of one of the rats sera taken at various time points against the T47D cell line. The counts are increasing with time (and subsequently with immunisations). The high background (Day -1) was within the range of all 12 unimmunised animals we tested (ranged from 25,000 to 40,000 cpm), and therefore we regarded it as normal.

Moreover we confirmed the above resultε with an ELISA using fixed T47D cells. The data is presented as area underneath the curve after diluting out the test sera (Table 2).

TABLE 2

DAYS Counts per min. Area

-1 37,559 0.098

12 100,887 0.161

22 150,975 0.188

26 188,215 0.392

+ve 0.515

TABLE 2: Rat anti-T47D response at various time points following immunisation with the HMFG2.

The results are presented as c.p.m. in an indirect radioimmunoaεsay, and as area underneath the curve in an

ELISA.

Production of anti-id² antibody:

Inhibition: Table 3A shows the properties of variouε supernatants in inhibiting binding of HMFG2 on the T47D cells using the Radioimmunoassay described (anti-mouse Ig iodine-125 labelled, as second layer). Competition: Table 3B shows the ability of the above tested supernatants to recognise the T47D Cells (anti-rat Ig iodine-125 labelled, as second layer).

TABLE 3A

Test Counts per min

PBS 10,000

HMFG2 25,856

Irrelevant 10,377 rat IgG

HMFG2 + irrelevant rat 30,532

IgG

GMN-B4+ 15,122

HMFG2

GMN-A1+ 20,932

HMFG2

GMN-A10+ 21,154

HMFG2

GMN-A9+ 20,193

HMFG2

TABLE 3A: Competition of HMFG2 binding to live T47D cells, in the presence of various rats supernatants (indirect radioimmunoassay) .

TABLE 3B

Test Counts per min.

PBS 4,983

Irrelevant rat IgG 9,454

GMN-B4 75,571

GMN-Al 21,887

GMN-A10 17,517

GMN-A9 3,845

TABLE 3B: Reactivity of various rat supernatents against live T47D cells (indirect radioimmunoassay).

Anti-MFG membrane extract activity:

When the above positive supernatant (GMN-B4) was tested against the purified mucin using an ELISA, it was found positive. Table 4 shows the data expresεed aε areas underneath the curve after several dilutions of both the test supernatant and the positive control (HMFG2). TABLE 4

TEST AREA

-ve 0.049

GMN-B4 0.323

HMFG2 0.475

TABLE 4: ELISA presented as area underneath the curve using the rat monoclonal antibody GMN-B4 against fixed T47D cells, the HMFG2 as positive control and and PBS as negative control.

Immunohistochemistry:

Figures 12 and 15 show positive immunoperoxide staining of the GMN-B4 supernatant on paraffin embedded tissues. Figures 13 and 16 are the negative controls and finally Figures 14 and 17 show the staining pattern of the HMFG2 on the same. Characterisation of the anti-id² antibody:

When tested the GMN-B4 was found to be a rat IgG2a.

Claims

1. A process for producing anti-idiotype² antibodies against an antigen which process comprises inoculating an animal with monoclonal antibodies, or fragments thereof, against the antigen and recovering antibodies against the antigen from a body fluid of the animal.

2. A process for producing cells capable of producing anti-idiotype² antibodies against an antigen which procesε compriεes inoculating an animal with monoclonal antibodies or fragments thereof, against the antigen, recovering from a tiεsue or body fluid of the animal cells capable of producing anti-idiotype² antibodies against the antigen and either (a) immortalising the cells or (b) recovering nucleic acid encoding the anti-idiotype² antibody from the cells and transfecting cells of an expression system with copies of the nucleic acid in expressible form.

3. A procesε for producing anti-idiotype² antibodies comprising either (a) culturing immortal or transfectant cells produced according to claim 2 and recovering antibodies from the culture medium or (b) injecting immortal or transfectant cells produced according to claim 2 into an anima-1 and recovering antibodies from the ascites fluid of the animal.

4. Anti-idiotype² antibodies produced according to the process of claim 1 or claim 3.

5. Immortal or transfectant cells capable of secreting anti-idiotype² antibodies and produced according to the process of claim 2.

6. Antibodies according to claim 4 for use in a method of treatment of the human or animal body by therapy or in a method of diagnosis practised on the human or animal body.

7. Use of antibodies according to claim 4 in the preparation of a medicament for use in a method of therapy or diagnosis practised on the human or animal body.

8. A diagnostic testing process which comprises contacting a sample suspected to contain an antigen with antibodies against the antigen according to claim 4.

9. A diagnostic test kit comprising antibodies according to claim 4.or cells according to claim 5.

10. A method of therapy or diagnosis comprising administering to a human or non-human animal in need thereof an effective non-toxic amount of an antibody according to claim 4.